The present invention relates to a process and an apparatus for real-time determination of solid sample composition through glow discharge optical emission spectroscopy (GD-OES) as a function of the sputtered depth in the sample and to an apparatus for implementing the process.
As is well known, glow discharge optical emission spectroscopy (GD-OES) is a rapid technique for surface or bulk analysis of solid materials.
GD-OES combines a glow discharge with an optical emission spectrometer.
The solid sample to be analysed constitutes the cathode of the glow discharge device. The material of the cathode is impacted by positive ions of an argon plasma and sputtered atoms from the cathode enter the plasma where they are excited by collisions with the more energetic electrons or by collisions with excited metastable argon atoms. These excited atoms are de-excited by optical emission, hence producing a xe2x80x9cglowxe2x80x9d. De-exciting atoms emit photons with characteristic wavelengths. By measuring the signals of these wavelengths, one can then measure the numbers of each type of atoms coming from the cathode and therefore the sample composition.
The crater depth formed by sputtering the sample atoms increases as the process goes on. Thus, the composition of the sample can be determined as a function of the time. These time dependant measurements can be converted in a quantitative result, i.e. a measurement as a function of the depth within the sample.
However, it necessitates a calibration and the use of calculation algorithms.
The conversion is made afterwards, i.e. after the measurement has been done.
More precisely, quantification can be carried out based on the following theoretical basis:
There are three primary processes in generation of the analytical signal:
1. the supply of sputtered atoms,
2. excitation followed by de-excitation, and
3. detection.
It is normally assumed these processes are independent.
Hence the recorded signal for a given emission line from element I is given by
li=ki.ei.qixe2x80x83xe2x80x83(1)
Where, from right to left, qi is the supply rate of element i into the plasma, ei represents the emission process, and ki is the instrumental detection efficiency.
The supply rate qi, which is also the elemental sputtering rate, will vary with the concentration, ci, of element i in the sample and with the overall sputtering rate, q, so that:
qi=ci.qxe2x80x83xe2x80x83(2)
The emission term will vary with the number of photons emitted per sputtered atom and with the absorption of these photons in traversing the plasma to reach the source window, so that:
ei=Si.Rixe2x80x83xe2x80x83(3)
where Ri is the emission yield and Si is a correction for self-absorption and will vary between 0 and 1 depending on the elemental sputtering rate. The detection efficiency is assumed constant. To these must be added a background term, bi, originating from photomultiplier dark current, instrument noise, scattered light, argon emission and unwanted signals from nearby emission lines.
These considerations lead to the following general equation for GD-OES:
xe2x80x83li=ki.Si.Ri.ci.q+bixe2x80x83xe2x80x83(4)
Equation (4) in fact represents a set of equations, one for each element i in the sample. If the terms ki, Si, Ri, and bi are constant then li will vary linearly with ciq.
In any analysis or at any particular depth in a depth profile, when the signals from all of the elements with significant concentrations are recorded, it can be assumed the concentrations will add up to 100%, i.e.:                                           ∑            i                    ⁢                      c            i                          =        1                            (        5        )            
When the set of equations represented by equation (4) is solved simultaneously with equation (5) the solution provides note only all the concentrations at the depth where the signals were recorded but also the instantaneous sputtering rate q at that depth.
Thus, the simultaneous solution to equations (4) and (5) provides first, a set of concentrations:
Ci=[(lixe2x88x92bi)/(ki.Ri.Si)]/qxe2x80x83xe2x80x83(6)
and, secondly, the sputtering rate:                     q        =                              ∑            i                    ⁢                                    (                                                l                  i                                -                                  b                  i                                            )                        /                          (                                                k                  i                                ·                                  R                  i                                ·                                  S                  i                                            )                                                          (        7        )            
To obtain a quantitative depth profile from equations (6) and (7), it is necessary to determine the concentration as a function of depth. To do this the value of q estimated at each depth is converted from xcexcg/s to xcexcm/s, i.e. to a penetration rate w, through some assumption about density, and then these penetration rates are integrated to determine the depth, z, i.e.
Z=xcexa3w(t).xcex94txe2x80x83xe2x80x83(8)
This quantification is long, difficult to implement and lacks of accuracy.
It necessitates the prior measurement of the erosion rates of calibrating samples. The measurement of these erosion rates requires the use of devices such as profilemeters and balance. The error in precision of the measurement is at best of 5% which adds to other errors. Furthermore, the conversion time/depth also depends upon an estimation of the sample density in function of the chemical composition: the algorithms use for the calculation do not take perfectly into account the nature of the sample (oxides, nitrites, etc).
In summary, the measurement error is at best about 10% for known samples and can increase up to 50% for unknown samples.
Thus, the aim of the present invention is to provide a process for real-time determination of a solid sample composition through glow discharge optical emission spectroscopy which overcomes the drawbacks of the prior art processes.
The above goal is achieved according to the invention by providing a process for real-time determination of a solid sample composition which comprises:
a) forming a glow discharge of atoms sputtered from an exposed area of the sample, and analysing the glow discharge by optical emission spectroscopy;
b) measuring the distance between said exposed area and a fixed reference surface and determining from this measured distance the depth of the exposed area within the sample; and
c) correlating the determined depth of the exposed area with the glow discharge analysis.
In a first implementation mode of the process of the invention, all process steps are carried out either continuously or periodically, so that there is obtained a continuous or periodical sample composition analysis as a function of the depth of the exposed area within the sample. In a second implementation mode of the process of the invention, measuring of the distance between the exposed area and the reference surface and determination of the exposed area depth is carried out at the end of the glow discharge analysing step, and the process further comprises the steps of:
d) calculating, after step (a), the theoretical depth of the exposed area within the sample using classical algorithms based on sample density estimation; and
e) comparing the determined depth and the calculated depth to decide whether or not there is a clear error in the estimated sample density as indicated by a significant difference between the determined and the calculated depths.
In a third implementation mode of the process, the process is used to analyse a multilayer sample where thicknesses of the layers are approximately known and each layer contains at least a known element which can be used as an indicator of the layer.
In that case, measuring of the distance between the exposed area and the reference surface is carried out each time the indicator is highlighted in the glow discharge (either appearing or disappearing), thereby obtaining a precise determination of the positions of the layers within the sample as well as an accurate measurement of their thicknesses.
Although any kind of glow discharge device can be used in the process of the invention, radio-frequency glow discharge device is preferably used since it allows analysis of both conductive and non-conductive samples.
Measurement of the distance between the exposed area of the sample and the reference surface can be done using confocal microscopy or interferometry. Preferably, this measurement is done using laser interferometry. These measurement methods are well known per se.
The present invention also relates to an apparatus for implementing the process according to the invention which comprises:
means for creating a glow discharge of atoms sputtered from an exposed area of a sample to be analysed;
means for spectroscopically analysing the glow discharge;
means for measuring the distance between said exposed area and a reference surface;
means for determining from the measured distance, the depth of the exposed area within the sample; and
means for correlating the glow discharge spectroscopic analysis with the determined depth of the exposed area.
In a preferred embodiment of the apparatus, the means for measuring the distance between the exposed area of the sample and the reference surface is a laser interferometer and the means for determining the depth of the exposed area within the sample and the correlation means are a suitably programmed computer.